The agricultural modelling world has generated several models aiming at the analysis of the response of the sector to certain changes in exogenous mainly policy variables. Among those, the CAPRI modelling system developed by a consortium centred on the University of Bonn and the AGLINK-COSIMO model, a joint product of the OECD and the FAO, are well known and accepted as comprehensive tools. This analysis focuses on a qualitative comparison of both models and particularly on the process of setting up the baseline. The baseline is a medium-term projection of agricultural markets reflecting current policies and those already decided upon. This projection in turn serves as the base for comparisons when analyzing scenarios. It is shown that CAPRI uses generic and automatic procedures whenever possible for conducting the database and the baseline, while AGLINK-COSIMO puts more emphasis on expert knowledge in this process. Both approaches are shown to have certain advantages while the conclusion that a combination of them would potentially improve both models will be drawn from this analysis.